AdPsyche

This is a repository for AdPsyche, Japanese preference dataset based on advertising psychology.

📁 Dataset Files

This repository contains two main datasets used in our study on human preferences in advertisement. The datasets are intended for analyzing the relationship between user personality traits and ad appeal.

---

human_preference.csv

This file contains pairwise comparison annotations of Japanese advertisements. Each row represents a human judgment about which of two ads is more appealing for the same product.

Columns:

Column Name	Description
product_id	Unique identifier for the product
lp_text	Landing page text (contextual product description shown to annotators)
ad_text_a	First ad text (option A)
ad_text_b	Second ad text (option B)
appeal_type_a	Type of appeal for ad A (e.g., "無料"=free, "高品質"=High quality, etc.)
appeal_type_b	Type of appeal for ad B
preferred_${user_id}	Individual judgments from up to 109 annotators for each ad pair (e.g., a. ad_text_a, b. ad_text_b, or same)

---

big5_zscores.csv

This file contains standardized Big Five personality trait scores (z-scores) for each annotator in the dataset. It enables trait-based filtering and analysis across the participant group.

Columns:

Column Name	Description
user_id	Unique identifier for the annotator (matches IDs used human_preference.csv)
gender	Annotator's gender (e.g., "男性" = male, "女性" = female)
age	Age group (e.g., "30代" = 30s, "40代" = 40s, etc.)
z_neuroticism	Z-score for Neuroticism
z_extraversion	Z-score for Extraversion
z_openness	Z-score for Openness
z_agreeableness	Z-score for Agreeableness
z_conscientiousness	Z-score for Conscientiousness

---

Citation

Thank you for your interest in our dataset. If you use it in your research, please cite:

@inproceedings{mita-etal-2025-adpsyche,
    title = "AdPsyche: 広告心理学に基づく選好データセット",
    author = "三田, 雅人  and
      村上, 聡一朗  and
      本多, 右京  and
      岡, 達志",
    booktitle = "言語処理学会 第31回年次大会 発表論文集",
    month = 3月,
    year = "2025",
    publisher = "言語処理学会",
    url = "https://www.anlp.jp/proceedings/annual_meeting/2025/pdf_dir/P10-13.pdf",
    pages = "4115--4120",
}

License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.